LETTERS TO THE EDITOR
Single-Nucleotide Polymorphism rs498055
on Chromosome 10q24 Is Not Associated
with Alzheimer Disease in Two Independent
Family Samples
To the Editor: In the January issue of the American Journal
of Human Genetics, Grupe and colleagues1published
evi-dence suggesting genetic association between SNP rs498055 on chromosome 10q24, located in a putative homologue of ribosomal protein S3a (RPS3A [MIM 180478]), and risk for Alzheimer disease (AD [MIM 104300]) in four of six independent case-control samples. The authors reached this conclusion after testing nearly 1,400 SNPs, using an exploratory case-control sample, followed by assessments of a number of independent data sets of different size, origin, and ascertainment. Although three of the replica-tion samples showed significant risk effects for the G allele of rs498055, this effect was not confirmed in two smaller series of neuropathologically confirmed AD cases and con-trols. None of the other 68 “hits” uncovered in the first pass received the same degree of consistent replication as did rs498055. Overall, the effect of the putative risk allele was modest (yielding odds ratios [ORs] between∼1.3 and 1.4) and—according to the authors’ conclusion—likely re-flects linkage disequilibrium (LD) with another genetic variant nearby. Here, we have set out to independently assess the association between rs498055 and AD risk in two large and carefully characterized samples of AD-af-fected families comprising nearly 1,900 subjects from 654 pedigrees. However, in contrast to the findings of Grupe and colleagues, we observed no evidence of association between rs498055 and AD in any of our analyses.
Using high-efficiency fluorescence polarization (HEFP) technology, we genotyped this SNP in two family-based AD samples: (1) 1,439 subjects from 437 multiplex AD-affected families recruited as part of the National Institute of Mental Health (NIMH) Genetics Initiative AD Study Sample (average age at onset [ⳲSD] of affected individuals was 72.4Ⳳ 7.7 years) and (2) 489 subjects from 217 in-dependent families, mostly consisting of discordant sib-ships, recruited as part of the Consortium on Alzheimer’s Genetics (CAG) (average age at onset was71.2Ⳳ 9.1years). These samples, as well as the genotyping procedures, are described in detail in the work of Bertram et al.2; PCR and
HEFP primer sequences for rs498055 are available on re-quest. Average genotyping efficiency across both samples was 98.4%, with a genotyping error rate!1% (on the basis
of∼10% duplicated samples). Power analyses (fig. 1) in the combined sample showed that, at a disease-allele fre-quency of 0.47 (i.e., the average frefre-quency of the G allele in U.S. controls reported by Grupe et al.) and a p.05, power was 64% for an OR of 1.3 and was 83% for an OR of 1.4 (see fig. 1 for more details). Naturally, power was
lower for the two samples considered separately, but it was still 40%–60% for the NIMH sample alone, comparable to the power of the replication samples in the study by Grupe et al.
In contrast to the findings of Grupe et al., we did not observe any significant evidence of association between
rs498055 and AD risk, neither in the two samples
indi-vidually nor after combining both data sets (table 1), over-all or when stratified by age at onset (with age 65 years as cutoff) or apolipoprotein E (APOE [MIM 107741]) 4-carrier status. Interestingly, and in contrast to the over-transmission of the G allele noted by Grupe et al., in our two samples, this allele was generally undertransmitted to affected individuals, which approached statistical sig-nificance in two of our stratified analyses in the combined sample (P p .09 in “late-onset” families, andP p .06 in “APOE 4-positive” families [table 1]). Finally, we also tested for association between rs498055 and age at onset of AD (age at last examination of unaffected individuals), using FBAT-LOGRANK, FBAT-Wilcoxon, and FBAT-Flem-ington-Harrington3,4in the unstratified samples. However,
none of these tests showed even marginally significant P values (data not shown). Since quantitative trait analyses are expected to be more powerful than analyses of binary traits if the underlying association is true,3these results
strengthen our overall negative conclusion.
Our study is the first to independently assess the poten-tial association between rs498055 and AD that had emerged from a semisystematic screen of 1,397 SNPs on chromo-some 10.1The fact that we failed to replicate the previous
findings is noteworthy for several reasons. First, rs498055 is located within the chromosome 10q24 linkage peak re-ported elsewhere for this collection of NIMH families.5
Thus, our sample should be particularly well suited to de-tect disease associations underlying this linkage signal. Second, our study is the first to analyze this SNP with use of family-based methodologies in which affected subjects are compared with related unaffected subjects from the same family. Results from such analyses are more robust to bias due to population admixture or other sources of skewed genotype distributions in cases or controls; this is of particular note, given the differences in allele frequen-cies reported for the two control populations from the St. Louis area in the work of Grupe et al. (see below). Despite the strengths of our approach, it is possible that we have missed a putative risk effect at rs498055 because of insuf-ficient power, especially when aiming to detect minor ef-fects with ORs of⭐1.3 (fig. 1). However, the differences between our findings and those of Grupe et al. are unlikely to result from lack of power alone, since we see under-rather than overtransmission of the G allele in both sam-ples. It is unclear whether these discrepancies are caused by chance or by differential patterns of LD across the var-ious samples.
Figure 1. Power to detect a range of effect sizes in the family samples analyzed. Power estimates were done with PBAT (v3.1).12
Estimates are based on approximation and are calculated for an additive disease model based on parameters published by Grupe et al.1
(i.e., disease-allele frequency of 0.47 and OR 1.2–1.6), with the exception of disease prevalence, which was set to 10%. Although the precise prevalence of AD is unknown and difficult to estimate, power does not change appreciably when prevalence is varied from 5% to 15% (data not shown). ORs are for heterozygous carriers of the disease allele versus homozygous noncarriers. Families were modeled after the observed pedigree structure for each sample, with both parents set as “missing.” Note that PBAT can currently handle a maximum offspring number of only four; however, 68 (16%) of NIMH pedigrees actually have more than four genotyped and phenotyped offspring, so that the power for “NIMH” and “NIMH⫹CAG” is likely to be underestimated (see the PBAT Web site for more details).
The long arm of chromosome 10 has been a focus of work for many AD genetics laboratories since the discovery of significant linkage with AD phenotypes by three inde-pendent groups, including ours (AD6 [MIM 605526]).5–7
These publications were followed by two additional stud-ies suggesting the presence of an AD risk and/or age-at-onset–modifying gene on this chromosome.8,9Although
nearly 30 positional candidate genes have since been as-sessed as potential AD risk factors underlying these linkage signals and several positive association findings have been published, no gene has received consistent support from independent follow-up studies,10and none shows evidence
of conclusive and significant summary effects in system-atic meta-analyses of all published and available genotype data (AlzGene).
Unfortunately, the present failure to replicate the prom-ising results of Grupe and colleagues is consistent with this overall pattern. There are several possible reasons for the differences between our findings and theirs. First, the difference might be due to chance, because the initial find-ing is a false-positive result. The 69 hits among 1,397 SNPs in the exploratory data set is close to the expected value by chance alone, as is the confirmation of 5 of these 69 signals in at least one of the two direct follow-up samples. However, we agree with the authors that a significant
over-representation of the same allele in three of five confir-mation samples, as observed for rs498055, may be unlikely to occur by chance alone. Second, the difference might have arisen by chance because our finding is a false-neg-ative result. Although this is possible, it should be noted that our sample is as large or larger than many replication samples in the field. In addition, the difference is unlikely to result from insufficient power alone, because the puta-tive risk allele, if anything, is undertransmitted in our sam-ples. Third, the differences may relate to our use of family-based methods, which are more robust to bias due to pop-ulation admixture. Although the degree to which admix-ture may lead to spurious association findings in case-control samples is controversial,11the issue is a concern
here, given the marked difference in allele frequencies across the two independent Washington University con-trol samples (47% for the case-concon-trol sample—similar to the other U.S. control sample—and 44% for the controls used in comparison with the linkage sample), a difference substantial enough that the allelic association between
rs498055 and AD in the linkage sample (49.8% risk-allele
frequency) would not have been significant had the other Washington University control set (or the University of California–San Diego controls) been used. Finally, the dif-ferences across studies may be due to difdif-ferences in
pat-Table 1. Association Analyses of rs498055 in Two Independent Family Samples
Samplea
FBAT Statistic Result G Allele Frequency No. of Informative Families Z Scoreb P All Families: NIMH .519 123 ⫺1.096 .27 CAG .504 84 ⫺.194 .85 Combined .514 207 ⫺1.001 .32
Families with late-onset disease:
NIMH .537 81 ⫺1.541 .12
CAG .531 66 ⫺.808 .42
Combined .535 147 ⫺1.703 .09
APOE4-positive families:
NIMH .494 107 ⫺1.488 .14
CAG .515 31 ⫺1.380 .17
Combined .499 138 ⫺1.929 .06
NOTE.—Association tests were performed using FBAT (v1.5.5) with an additive transmission model, the empirical variance function, and an equal-weight offset correction for affected and unaffected individuals (see the FBAT Web site for more details).
a Families were classified as “late onset” when all sampled affected individuals had age at onset of165 years and were classified as “APOE4 positive” when at least one affected individual per family carried the4 allele. The smaller strata of remaining families (i.e., those displaying an earlier age at onset or those in which none of the affected individuals carried an4 allele) also failed to show evidence of significant association (data not shown).
b For the G allele of rs498055, which was reported as the putative risk allele by Grupe et al.1 (positive values indicate overtransmission to affected individuals). Note that the direction of transmission is consistent for both family samples analyzed here and is opposite to that seen in the previous publication.1
terns of LD across the various samples, which are impos-sible to assess as long as the precise nature of the putative risk allele at this locus remains unknown.
Clearly, additional analyses of sufficiently powered and independent samples are needed to assert whether
rs498055, or a polymorphism in LD with it, makes a
rel-evant contribution to AD risk. At least in the two family samples investigated here—one of which shows linkage to the same chromosomal interval as rs498055—this SNP is not a major determinant of AD risk.
LARS BERTRAM, MONICA HSIAO, CHRISTOPH LANGE, DEBORAH BLACKER, AND RUDOLPH E. TANZI Acknowledgments
The authors thank all families for participating in this study. This work was sponsored by grants from the NIMH, the National Institute on Aging (NIA) (at the Alzheimer Disease Research Center), and the Alzheimer Association. L.B. is supported by NIA grant 1R01 AG023667-01 and by a Young Investigator Award from the National Alliance for Research on Schizophrenia and Depression.
Web Resources
The URLs for data presented herein are as follows:
AlzGene, http://www.alzgene.org/ (Bertram L, McQueen MB, Mul-lin K, Blacker D, Tanzi R. The AlzGene Database–Published AD Candidate Genes; Alzheimer Research Forum; last accessed 3/ 1/06)
FBAT, http://www.biostat.harvard.edu/˜fbat/fbat.htm
Online Mendelian Inheritance in Man (OMIM), http://www.ncbi .nlm.nih.gov/Omim/ (for RPS3A, AD, APOE, and AD6) PBAT, http://www.biostat.harvard.edu/˜clange/default.htm
References
1. Grupe A, Li Y, Rowland C, Nowotny P, Hinrichs AL, Smemo S, Kauwe JSK, et al (2006) A scan of chromosome 10 identifies a novel locus showing strong association with late-onset Alz-heimer disease. Am J Hum Genet 78:78–88
2. Bertram L, Hiltunen M, Parkinson M, Ingelsson M, Lange C, Ramasamy K, Mullin K, Menon R, Sampson AJ, Hsiao MY, Elliott KJ, Velicelebi G, Moscarillo T, Hyman BT, Wagner SL, Becker KD, Blacker D, Tanzi RE (2005) Family-based associ-ation between Alzheimer’s disease and variants in UBQLN1. N Engl J Med 352:884–894
3. Lange C, DeMeo DL, Laird NM (2002) Power and design con-siderations for a general class of family-based association tests: quantitative traits. Am J Hum Genet 71:1330–1341 4. Jiang H, Harrington D, Raby BA, Bertram L, Blacker D, Weiss
ST, Lange C (2006) Family-based association test for time-to-onset data with time-dependent differences between the haz-ard functions. Genet Epidemiol 30:124–132
5. Bertram L, Blacker D, Mullin K, Keeney D, Jones J, Basu S, Yhu S, McInnis MG, Go RC, Vekrellis K, Selkoe DJ, Saunders AJ, Tanzi RE (2000) Evidence for genetic linkage of Alzhei-mer’s disease to chromosome 10q. Science 290:2302–2303 6. Ertekin-Taner N, Graff-Radford N, Younkin LH, Eckman C,
Baker M, Adamson J, Ronald J, Blangero J, Hutton M, Younkin SG (2000) Linkage of plasma Ab42 to a quantitative locus on chromosome 10 in late-onset Alzheimer’s disease pedigrees. Science 290:2303–2304
Table 1. Frequency of AD Risk Alleles Is Higher in Unaffected Individuals from Multiply Affected Families Than in Unrelated Controls
Locus and Allele
No. (%) of Subjects Cases Controls Linkage CC Linkage CC APOE: 2 7 (2.3) 80 (4.2) 16 (5.2) 181 (8.5) 3 169 (54.9) 1,144 (60.2) 197 (64.4) 1,691 (79) 4 (Risk allele) 132 (42.8) 676 (35.6) 93 (30.4) 268 (12.5) rs498055: A 133 (46.2) 915 (47.9) 140 (46.4) 1,174 (54.4) G (Risk allele) 155 (53.8) 995 (52.1) 162 (53.6) 984 (45.6) NOTE.—CCp case-control study.
7. Myers A, Holmans P, Marshall H, Kwon J, Meyer D, Ramic D, Shears S, Booth J, DeVrieze FW, Crook R, Hamshere M, Abraham R, Tunstall N, Rice F, Carty S, Lillystone S, Kehoe P, Rudrasingham V, Jones L, Lovestone S, Perez-Tur J, Williams J, Owen MJ, Hardy J, Goate AM (2000) Susceptibility locus for Alzheimer’s disease on chromosome 10. Science 290:2304– 2305
8. Li Y-J, Scott WK, Hedges DJ, Zhang F, Gaskell PC, Nance MA, Watts RL, et al (2002) Age at onset in two common neuro-degenerative diseases is genetically controlled. Am J Hum Ge-net 70:985–993
9. Lee JH, Mayeux R, Mayo D, Mo J, Santana V, Williamson J, Flaquer A, Ciappa A, Rondon H, Estevez P, Lantigua R, Ka-warai T, Toulina A, Medrano M, Torres M, Stern Y, Tycko B, Rogaeva E, St George-Hyslop P, Knowles JA (2004) Fine map-ping of 10q and 18q for familial Alzheimer’s disease in Ca-ribbean Hispanics. Mol Psychiatry 9:1042–1051
10. Bertram L, Tanzi RE (2004) Alzheimer’s disease: one disorder, too many genes? Hum Mol Genet Spec No 1 13:R135–R141 11. Hoggart CJ, Parra EJ, Shriver MD, Bonilla C, Kittles RA, Clay-ton DG, McKeigue PM (2003) Control of confounding of ge-netic associations in stratified populations. Am J Hum Genet 72:1492–1504
12. Lange C, DeMeo D, Silverman EK, Weiss ST, Laird NM (2004) PBAT: tools for family-based association studies. Am J Hum Genet 74:367–369
From the Genetics and Aging Research Unit, MassGeneral Institute for Neurodegenerative Diseases, Department of Neurology (L.B.; M.H.; R.E.T.), and Gerontology Research Unit, Department of Psychiatry (D.B.), Mass-achusetts General Hospital, Harvard Medical School, Charlestown, MA; and Departments of Biostatistics (C.L.) and Epidemiology (D.B.), Harvard School of Public Health, Boston
Address for correspondence and reprints: Dr. Lars Bertram, Genetics and Aging Research Unit, MGH-East, MassGeneral Institute for Neuro-degenerative Diseases, 114 16th Street, Charlestown, MA 02129. E-mail: [email protected]
Am. J. Hum. Genet. 2006;79:180–183.䉷 2006 by The American Society of Human Genetics. All rights reserved.
0002-9297/2006/7901-0022$15.00
Reply to Bertram et al.
To the Editor: The study by Bertram and colleagues (in this issue)1failed to replicate, in two family-based sample
sets, the association of rs498055 with Alzheimer disease (AD [MIM 104300]) that we observed in four large, well-characterized case-control sample sets.2Although the
re-sult is disappointing, there are several differences between the studies that may have contributed to these discrepant findings. First, there are significant differences in the study designs. Bertram et al. used two family-based sample sets that included subjects with both early- and late-onset AD (e.g., 320 families with late-onset AD and 117 families with early/mixed-onset in the National Institute of Mental Health [NIMH] sample set) of different ethnicities (94% white; 6% others),3 which resulted in 147 informative
families with late-onset AD for both sample sets com-bined. The characterization of their unaffected controls
was based on self-assessment or a telephone interview, a procedure sufficient when “unaffecteds” are used solely to determine phase in linkage studies, but likely to sig-nificantly impact power in association studies, especially when familial loading is high, as it is in the sample of Bertram et al. Indeed, the authors acknowledge this in one of their previous publications by pointing out that the characterization of controls “may miss some mild cases of dementia” and lead “to a decrease in power.”3In contrast,
our study included only clinically evaluated, late-onset cases and nondemented controls of white origin. Second, the use of a family-based sample that was ascertained on the basis of multiple affected relatives is likely to partic-ularly adversely impact power to detect a risk allele of relatively high frequency and small effect size, such as
rs498055. Under these circumstances, the allele frequency
in unaffected relatives also increases,4with consequent
loss of power in comparison with case-control studies such as our own. To investigate this more fully, we compared the allele frequencies for a known genetic risk factor for AD, apolipoprotein E (APOE [MIM 107741]), and for the putative risk factor under debate, rs498055, in our com-bined case-control series and in the NIMH linkage families used by us in the study described by Myers et al.5In this
context, it is worth noting that 355 of 372 individuals from the linkage sample–derived cases in our recent pub-lication overlap with affected individuals in the NIMH family sample set described by Bertram et al. For the com-parison, we identified the subgroup of NIMH families with genotypes for at least one unaffected and one affected individual and then selected at random one unaffected and one affected individual from each of these families. Table 1 illustrates clearly that the frequency of the APOE4 allele is substantially higher in unaffected individuals from the linkage families than in unaffected individuals from the case-control series (30.4% vs. 12.5%) and that, although the APOE4 allele frequency is highest in the linkage cases, the difference between the unrelated cases and controls is much greater than that between familial cases and re-lated controls (35.6% vs. 12.5% compared with 42.8% vs. 30.4%). As a result, the odds ratio (OR) for the APOE4 allele in the case-control series is 3.8, compared with only 1.7
Table 1. LD between SERPINE2 SNPs Expressed as r2 SNP r2for SNP rs1438831 rs920251 rs6747096a rs3795879 rs920251 .952 (1.0) … … … rs6747096a .140 .148 … … rs3795879a .140 (.145) .145 (.145) .964 … ss49785625a .020 .023 .054 .055 NOTE.—The r2values in parentheses are values obtained from HapMap and compared with our own data in controls. ss49785625 and rs6747096 are not in HapMap.
a SNP reported by DeMeo et al.1to be associated with disease in both family and case-control cohorts.
for the family-based samples. Similarly for rs498055, the difference in frequencies between the cases and controls is greater for the unrelated samples than for the linkage families (table 1). Thus, the failure of Bertram et al. to replicate our results does not necessarily indicate that the original association was a false-positive result. We concur with Bertram et al. that the significant association of
rs498055 in four of six samples “may be unlikely to occur
by chance”1(p181)(in this issue). However, it is possible that
our initial study provided an overestimate of the allelic OR for rs498055. If this were true and the OR were!1.3,
then the study by Bertram et al. would clearly be under-powered. Further replication in well-characterized sample sets is required to assess whether the association is gen-uine. Ideally, this should be done with large case-control sample sets, to achieve maximum power. For this partic-ular marker, we estimate that 360 cases and 360 controls are needed to achieve 80% power in a replication study (one-sided a p.05), assuming an allelic OR of 1.3 and a risk-allele frequency of 45.6%. A meta-analysis of all studies should then be performed to determine whether
rs498055 is associated with late-onset AD. In addition, it
might be interesting to test the other reported significant markers from this region in additional sample sets.
ANDREW GRUPE, YONGHONG LI, CHARLES ROWLAND,
TONY HINRICHS, PETER HOLMANS, JOHN HARDY,
MICHAEL O’DONOVAN, MICHAEL J. OWEN,
JULIE WILLIAMS, AND ALISON GOATE Acknowledgments
Funding for this work was partly provided by National In-stitutes of Health grant RO1 AG16208 (to A. Goate), the Medi-cal Research Council, UK (to J.W., M.J.O., and M.O.), and the Alzheimer’s Research Trust (to J.W. and M.J.O). J.H. is supported by the National Institutes of Health intramural program and by the VERUM Foundation (DIADEM project).
Web Resource
The URL for data presented herein is as follows:
Online Mendelian Inheritance in Man (OMIM), http://www.ncbi .nlm.nih.gov/Omim/ (for AD and APOE)
References
1. Bertram L, Hsiao M, Lange C, Blacker D, Tanzi RE (2006) Single-nucleotide polymorphism rs498055 on chromosome 10q24 is not associated with Alzheimer disease in two independent family samples. Am J Hum Genet 79:180–183 (in this issue) 2. Grupe A, Li Y, Rowland C, Nowotny P, Hinrichs AL, Smemo
S, Kauwe JSK, et al (2006) A scan of chromosome 10 identifies a novel locus showing strong association with late-onset Alz-heimer disease. Am J Hum Genet 78:78–88
3. Bertram L, Hiltunen M, Parkinson M, Ingelsson M, Lange C, Ramasamy K, Mullin K, Menon R, Sampson AJ, Hsiao MY, El-liott KJ, Velicelebi G, Moscarillo T, Hyman BT, Wagner SL, Becker KD, Blacker D, Tanzi RE (2005) Family-based association between Alzheimer’s disease and variants in UBQLN1. N Engl J Med 352:884–894
4. Risch N, Teng J (1998) The relative power of family-based and case-control designs for linkage disequilibrium studies of com-plex human diseases I: DNA pooling. Genome Res 8:1273–1288 5. Myers A, Wavrant De Frieze F, Holmans P, Hamshere M, Crook R, Compton D, Marshall H, et al (2002) A full genome screen for Alzheimer’s disease: stage two analysis. Neuropsychiatric Genetics 114:235–244
From Celera Diagnostics, Alameda, CA (A. Grupe; Y.L.; C.R.); Depart-ments of Psychiatry, Neurology, and Genetics, Washington University, St. Louis (T.H.; A. Goate); Biostatistics and Bioinformatics Unit and Depart-ment of Psychological Medicine, Wales College of Medicine, Cardiff Uni-versity, Cardiff (P.H.; M.O.; M.J.O.; J.W.); and National Institute on Aging, Bethesda (J.H.)
Address for correspondence and reprints: Dr. Alison Goate, Department of Psychiatry, B8134, Washington University School of Medicine, 660 S. Euclid Avenue, St. Louis, MO 63110. E-mail: [email protected]
Am. J. Hum. Genet. 2006;79:183–184.䉷 2006 by The American Society of Human Genetics. All rights reserved.
0002-9297/2006/7901-0023$15.00
The SERPINE2 Gene and Chronic Obstructive
Pulmonary Disease
To the Editor: In the February 2006 issue of the Journal,
DeMeo et al.1identified SERPINE2 as a positional
candi-date gene for susceptibility to chronic obstructive pul-monary disease (COPD [MIM 606963]) and reported on the association of polymorphic variants of this gene with early-onset disease in a family-based study and with severe disease in a case-control study. With early prior infor-mation provided by the authors, we have independently tested for an association of the SERPINE2 gene with COPD in the largest case-control study reported to date. Our study consists of 1,018 COPD cases and 911 controls pro-spectively recruited from six European centers. We have provided details about the patients elsewhere.2The study
population was screened for genotypes at the Medical Re-search Council (United Kingdom) Gene Services Unit for five SNPs (table 1) in the SERPINE2 gene. All the SNPs evaluated were reported in the study by DeMeo et al. as associated with disease, with three of the five associated with disease in both the family and case-control study cohorts they assessed.
Table 2. SERPINE2 Genotype and Allele
Frequencies in Controls and COPD Cases
SNP and Sample Frequency of Allele Frequency of Genotype C T CC CT TT rs1438831: COPD case .66 .34 .43 .45 .12 Control .66 .34 .43 .46 .11 A G AA AG GG rs920251: COPD case .35 .65 .13 .45 .42 Control .35 .65 .12 .46 .42 A G AA AG GG rs6747096: COPD case .79 .21 .61 .35 .04 Control .79 .21 .63 .33 .04 C T CC CT TT rs3795879: COPD case .78 .22 .60 .36 .05 Control .79 .21 .62 .33 .05 A G AA AG GG ss49785625: COPD case .54 .46 .30 .48 .22 Control .53 .47 .29 .48 .23
We examined linkage disequilibrium (LD) between the SNPs (table 1) and evaluated SNP and haplotype associa-tions as described elsewhere.2DeMeo et al. did not report
specific LD values between SNPs or noncontiguous SNPs contributing to haplotypes. SNPs and genotype frequen-cies in the study population are shown in table 2. We found no significant deviation from Hardy-Weinberg equi-librium in frequencies for any of the SNPs.
We found no association between any of the SERPINE2 SNPs and disease, in examining both the allelic and ge-notype distributions, although our study was well pow-ered to detect associations of the magnitude observed by DeMeo et al., and we would have expected to see these frequency differences with the SNPs that we studied. We also failed to find a relationship between any haplotypes of these SNPs and disease (data not shown). It was of in-terest that the allele and genotype frequencies observed in our control and patient groups were virtually identical to those observed in control subjects by DeMeo et al., indicating a common distribution of SERPINE2 variants in the European and North American populations studied. Our previous study has also shown that there is no evi-dence of population stratification in our sample.
Patients evaluated in both the family-based and case-control studies reported by DeMeo et al. represent a severe subset of the disease spectrum. To determine whether the association with SERPINE2 noted by DeMeo et al. was re-lated to disease severity, we also analyzed SNP allele and genotype frequencies in the subgroup of our patients with forced expiratory volume at 1 s⭐45% (n p 388), a group that represents severe disease, but we failed to observe any association.
Our inability to replicate the observations of DeMeo et al. in a more highly powered case-control study may be related to differences in the disease phenotype of the pa-tients studied, because our papa-tients included those with and without emphysema. The possibility, however, that the associations reported by DeMeo et al. represent false-positive results must also be considered. In this respect, it is of note that, in the study by DeMeo et al., different associations were reported for SNPs that are in linkage disequilibrium with one another. For example, rs3795879 and rs3795877 have an r2 value of 1 in HapMap, yet
different associations with quantitative spirometric phe-notypes were reported for the family study. Similarly,
rs1438831 and rs920251 are in complete LD, with an r2
value of 1 in HapMap and 0.95 in our study; however, in DeMeo et al.’s case-control study, the allele and genotype frequencies of rs920251 were found to be significantly as-sociated with disease (P values of 0.015 and 0.011, re-spectively), whereas no similar association was observed for rs1438831. In both instances, the almost complete linkage between these pairs of SNPs would be expected to result in similar associations.
These results underline the importance of replication in other large independent studies before SERPINE2 can be unequivocally assigned as a candidate gene for COPD. It
is becoming apparent that, to detect modest genetic effects for complex diseases, several independent studies may be required and the data may need to be subjected to meta-analysis. For example, this approach has been used to study Alzheimer disease (see Alzheimer’s Association Web site). Similar approaches need to be adopted for COPD. It would also be helpful to have similar criteria adapted for phenotypic selection and to plan prospective studies on this basis.
SALLY CHAPPELL, LESLIE DALY, KEVIN MORGAN,
TAMAR GUETTA BARANES, JOSEP ROCA, ROBERTO RABINOVICH, ANN MILLAR, SEAMAS C. DONNELLY, VERA KEATINGS, WILLIAM MACNEE, JAN STOLK, PIETER S. HIEMSTRA, MASSIMO MINIATI, SIMONETTA MONTI, CLARE M. O’CONNOR,* AND NOOR KALSHEKER*
Acknowledgments
We thank Ed Silverman and Dawn DeMeo for providing us with detailed information about the SERPINE2 SNPs. This work was supported by European Union 5th Framework Programme grant QLG1-CT-2001-01012.
Web Resources
The URLs for data presented herein are as follows: Alzheimer’s Association, http://www.alz.org/
Online Mendelian Inheritance in Man (OMIM), http://www.ncbi .nlm.nih.gov/Omim/ (for COPD)
References
1. DeMeo DL, Mariani TJ, Lange C, Srisuma S, Litonjua AA, Ce-ledon JC, Lake SL, Reilly JJ, Chapman HA, Mecham BH, Haley KJ, Sylvia JS, Sparrow D, Spira AE, Beane J, Pino-Plata V, Speizer FE, Shapiro SD, Weiss ST, Silverman EK (2006) The SERPINE2 gene is associated with chronic obstructive pulmonary disease. Am J Hum Genet 78:253–264
2. Chappell S, Daly L, Morgan K, Guetta Baranes T, Roca J, Ra-binovich R, Millar A, Donnelly S, Keatings V, MacNee W, Stolk J, Hiemstra P, Miniati M, Monti S, O’Connor CM, Kalsheker N (2006) Cryptic haplotypes of SERPINA1 confer susceptibility to chronic obstructive pulmonary disease. Hum Mutat 27:103– 109
From the University of Nottingham and Division of Clinical Chemistry, Molecular Medical Sciences, Institute of Genetics, University Hospital, Queens Medical Centre, Nottingham, United Kingdom (S.C.; K.M.; T.G.B.; N.K.); Departments of Public Health Medicine and Epidemiology (L.D.) and Medicine and Therapeutics (S.C.D; C.M.O.), The Conway Institute, University College Dublin, Dublin; Service de Pneumologia, Hospital Clin-ico y Provincial de Barcelona, Barcelona (J.R.; R.R.); University of Bristol and Lung Research Group, Department of Clinical Science at North Bristol, Southmead Hospital, Bristol, United Kingdom (A.M.); Letterkenny General Hospital, Letterkenny, Ireland (V.K.); University of Edinburgh, Respiratory Medicine, ELEGI Colt Laboratories, Edinburgh (W.M.); Department of Pul-monology, Leiden University Medical Center, Leiden, The Netherlands (J.S.; P.S.H.); and CNR Institute of Clinical Physiology, Pisa, Italy (M.M.; S.M.)
Address for correspondence and reprints: Dr. Noor Kalsheker, The Uni-versity of Nottingham, Division of Clinical Chemistry, Institute of Genet-ics, Queens Medical Centre, Nottingham, NG7 2UH, United Kingdom. E-mail: [email protected]
Am. J. Hum. Genet. 2006;79:184–186.䉷 2006 by The American Society of Human Genetics. All rights reserved.
0002-9297/2006/7901-0024X$15.00
* These two authors contributed equally to this work.
Reply to Chappell et al.
To The Editor: We appreciate the efforts of Chappell and
colleagues1to replicate our SERPINE2 findings. We
iden-tified SERPINE2 as a candidate gene for chronic obstructive pulmonary disease (COPD [MIM 606963]) on the basis of our gene-expression results (in both murine and human lung) and our genetic association analysis results in two study populations. Chappell et al. found no evidence for association of five SERPINE2 SNPs with COPD in their case-control study. As in many complex-disease genetic asso-ciation studies in general, and in previous COPD ge-netic association studies in particular,2the results are
inconsistent.
There are many potential explanations for these incon-sistent results, including population stratification, genetic heterogeneity, false-positive and/or false-negative results, differences in the number of SNPs genotyped, and phe-notypic heterogeneity.3 In comparing the results of our
two research groups for association analysis of SERPINE2 SNPs with COPD, phenotypic heterogeneity is of partic-ular importance. COPD is a syndrome composed of both
emphysema and airway disease, with variable contribu-tions of these processes in different individuals with COPD. Review of chest CT scans of probands from the Boston Early-Onset COPD Study—the population in which we performed family-based association analysis of COPD-related phenotypes—revealed that the vast major-ity of these probands had emphysema.4 Moreover, the
COPD cases in our case-control replication population were clearly selected for emphysema as part of the Na-tional Emphysema Treatment Trial (NETT). In addition, the Boston Early-Onset COPD Study probands and the NETT cases had very severe COPD. Thus, our test and replication populations were severely affected with COPD, typically with a substantial degree of emphysema. As noted by Chappell et al., our cases represent “a severe subset of the disease spectrum,”1(p185)and their cases
rep-resent a broader spectrum of severity, including individ-uals with and without emphysema. The differences in disease severity and emphysema may be important con-tributors to their nonreplication of our association find-ings. Also of note, although Chappell et al. genotyped five SNPs in SERPINE2, they did not genotype several other SNPs for which we observed replicated associations and LOD score reduction in conditional linkage models.
Chappell et al. also comment about apparently incon-sistent association results in our family-based and case-control association analyses among SNPs in tight linkage disequilibrium (LD). Modest differences in the statistical significance of the association analysis results were noted for several SNPs that are in strong but not complete LD in our study populations. There are reasonable explana-tions for these modest differences. (1) The SNP pairs men-tioned are not in complete LD; in our combined case-control cohort, the r2values were 0.93 for rs3795879 and rs3795877 and 0.91 for rs1438831 and rs920251. (2)
De-spite excellent genotype completion rates, there were slight differences in missing data between these SNP pairs. Of note, these were not the only SERPINE2 SNPs signifi-cantly associated with COPD-related phenotypes in our study; we observed 18 significantly associated SERPINE2 SNPs in the family-based association analysis and 7 sig-nificantly associated SNPs in the case-control analysis.
We fully agree with Chappell et al. that replication of significant associations is essential—which is why we in-cluded in our article the replication of our family-based association analysis results in a separate case-control study. This is also the reason why we provided early access to significantly associated SNPs to the Chappell and Kal-sheker group.
Is SERPINE2 a confirmed COPD susceptibility gene? Cer-tainly not. Before the impact of SERPINE2 on COPD sus-ceptibility is fully known, more genetic association studies as well as functional studies will be needed. However, we contend that SERPINE2 remains a valid COPD candidate gene. Finally, we agree with Chappell et al. that agreement on phenotypic definitions and collaboration between
re-search groups are crucial for the future of genetic studies of COPD and other complex diseases.
Web Resources
The URL for data presented herein is as follows:
Online Mendelian Inheritance in Man (OMIM), http://www.ncbi .nlm.nih.gov/Omim/ (for COPD)
References
1. Chappell S, Daly L, Morgan K, Guetta Baranes T, Roca J, Ra-binovich R, Millar A, Donnelly SC, Keatings V, MacNee W, Stolk J, Hiemstra PS, Miniati M, Simonetta M, O’Connor CM, Kalsheker N (2006) The SERPINE2 gene and chronic obstructive pulmonary disease. Am J Hum Genet 79:184–185 (in this issue) 2. Hersh CP, DeMeo DL, Lange C, Litonjua AA, Reilly JJ, Kwiat-kowski D, Laird N, Sylvia JS, Sparrow D, Speizer FE, Weiss ST, Silverman EK (2005) Attempted replication of reported chronic
obstructive pulmonary disease candidate gene associations. Am J Respir Cell Mol Biol 33:71–78
3. Silverman EK, Palmer LJ (2000) Case-control association stud-ies for the genetics of complex respiratory diseases. Am J Respir Cell Mol Biol 22:645–648
4. Silverman EK, Chapman HA, Drazen JM, Weiss ST, Rosner B, Campbell EJ, O’Donnell WJ, Reilly JJ, Ginns L, Mentzer S, Wain J, Speizer FE (1998) Genetic epidemiology of severe, early-onset chronic obstructive pulmonary disease: risk to relatives for air-flow obstruction and chronic bronchitis. Am J Respir Crit Care Med 157:1770–1778
From the Channing Laboratory and Division of Pulmonary and Critical Care Medicine, Department of Medicine, Brigham and Women’s Hospital, and Harvard Medical School, Boston (D.L.D.; E.K.S)
Address for correspondence and reprints: Dr. Dawn L. DeMeo, Channing Laboratory, 181 Longwood Avenue, Boston, MA 02115. E-mail: redld@ channing.harvard.edu
Am. J. Hum. Genet. 2006;79:186–187.䉷 2006 by The American Society of Human Genetics. All rights reserved.
